How sure are you that FM2 shipped with no MSAA? I could be mistaken, but I thought it shipped with variable rate (announced in a Japanese article) between 1x (none) to 4x, later being patched upward.
- sigh- they're all from Prologue (u can see the Suzuka track in the first 3 pics).
And they're all in game, except maybe the last one, the other all show a full HUD.
Upwards of 4x? Are there any other 360 games that use more than 4xMSAA?
yes i know they are, and you first were talking about GT5 and not GT5p.
and this comparison was pretty good BUT this was comparison was started by judging a print-screen of a none finished game.
and the poly count is pretty high up there, but as far as textures are concerned you're not going to really see them crisp looking through a picture of a video. are you?
comparisons although they are fun to do we can't do them just yet.
I'm pretty certain those are in game models.
Don't see any reason why they can't be 50k poly cars, probably less though (I'd guess those are 20-30k models). A quick rough count would put each wheel at around 2k (counting segments etc). As mentioned in the interview, poly counts really don't matter too much, it's all fillrate and what you do with the pixels.
For comparison, the stanford bunny is a good example, it's 68k polygons, (wireframe).
I managed to get it rendering at 1500fps, which is around about 105 million/tris/sec. (x1800xl) So you can see it's no trouble really. (pic). 12 50k poly cars at 60fps is 36 million tris/sec, making it quite possible. I've had 20 bunnies rendering full screen at over 60fps as well.
Upwards of 4x? Are there any other 360 games that use more than 4xMSAA?
Didn't the devs say "dynamic MSAA" from 0x - 4xAA ingame depending on the load? Looks more like 0xAA 90% of the time to me...
that was not final/inaccurate info.
final was 2xAA in game.... 4xAA replay... which is what the demo is. so you must not be seeing it right.
How sure are you that FM2 shipped with no MSAA? I could be mistaken, but I thought it shipped with variable rate (announced in a Japanese article) between 1x (none) to 4x, later being patched upward.
I seem to recall that when the demo first came out it didn't have antialiasing; shortly after it was added into it via a patch. When the game launched a bit later on, it had the variable antialiasing from the start.Not 100% sure. I might even be confusing the demo with the full game at this point. So you may be right. I did get the game and the demo at the moment it came out in Europe though, but it's been a while.
final was 2xAA in game.... 4xAA replay... which is what the demo is. so you must not be seeing it right.
the pictures of the rabbit really wouldn't apply to forza's cars cause of two reasons,
1, the rabbit's exterior was modeled, there was no interior modeling inside of the rabbit. (so naturally all of the polygons go to the exterior shell))
2, the rabit was done in triangles......(it takes two triangles to make a polygon, twice the mess)
Good to see that the AA question has been sorted, I was sure Forza 2 doesn't have variable AA in-game.
If Forza 3's in-game rendering res/AA is unchanged from F2, how come the car models look so much cleaner? In those recent vids the car models had less aliasing that even PGR4, and that used 4x AA
I think AlStrong is asking Che over on GAF if T10 is using the tesselator for Forza 3, which would shed more light on the 10x polys question.
Btw, as nice as the game is shaping up it is really clear that this generation of consoles has a lot to work on in regards to IQ and LOD. The photomode really shows how the assets are great but the realtime rendering can use some more umph.